Hi, I have Jetpack 4.2 and installed the tensorflow version 2.0 wheel from this index:
https://developer.download.nvidia.com/compute/redist/jp/v42/tensorflow-gpu/
following this guide:
but the CuDNN version that the package is compiled with (7.5) does not match the CuDNN version installed on this version of Jetpack (which is 7.3)
so i get the following error when running a simple CNN:
2020-05-03 13:29:40.416528: I tensorflow/stream_executor/platform/default/dso_loader.cc:44] Successfully opened dynamic library libcudnn.so.7
2020-05-03 13:29:42.356460: E tensorflow/stream_executor/cuda/cuda_dnn.cc:319] Loaded runtime CuDNN library: 7.3.1 but source was compiled with: 7.5.0. CuDNN library major and minor version needs to match or have higher minor version in case of CuDNN 7.0 or later version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.
2020-05-03 13:29:42.361580: E tensorflow/stream_executor/cuda/cuda_dnn.cc:319] Loaded runtime CuDNN library: 7.3.1 but source was compiled with: 7.5.0. CuDNN library major and minor version needs to match or have higher minor version in case of CuDNN 7.0 or later version. If using a binary install, upgrade your CuDNN library. If building from sources, make sure the library loaded at runtime is compatible with the version specified during compile configuration.
2020-05-03 13:29:42.363119: W tensorflow/core/common_runtime/base_collective_executor.cc:216] BaseCollectiveExecutor::StartAbort Unknown: Failed to get convolution algorithm. This is probably because cuDNN failed to initialize, so try looking to see if a warning log message was printed above.